Cluster Analysis for Long-Term Power Quality Data in Mining Electrical Power Network

This paper presents the results of using data mining for analysis of long-term power quality (PQ) data recorded in a supply power network of the mining industry. The analyzed network is characterized by prominent load variation caused by welding machines, conveyor belts and drainage pumps. Additionally, a distributed generation (DG, DER) is installed in the network represented by combined heat and power plant (CHP) and steam-gas generation units. In order to explore the PQ data, a cluster analysis (CA) is proposed a prominent representative of data mining (DM). Obtained results allow to indicate CA as a proper method for automatic identification of voltage event that is valuable for flagging concept as well as identification of periods of time when the network reveals different working conditions. It allows finally to investigate and compare the influence of distributed generation on electric power network of the mining industry.